STTR Phase I: A Reinforcement Learning-Based Automated Self-Guiding Drilling Tool

这项小型企业技术转让计划第一阶段（STTR）项目的更广泛影响/商业潜力是通过改善钻井过程中的信息流，提供一种环保、可持续和具有成本效益的能源获取途径。到2026年，美国地热能产业预计将增长48%，达到68亿美元，并将受益于钻井活动的显著节省，目前钻井活动占整个地热发电厂和油田开发成本的40-60%。此外，石油和天然气钻井作业将受益于更快、更清洁的作业。STTR一期项目解决了无法将大量数据从井下传输到地面的问题，主要是在钻头上进行处理。所提出的最先进的钻井工具结合了先进的处理能力和井下智能。它建立在以强化学习（RL）技术为中心的基于物理的机器学习算法之上，这种算法以前用于自动驾驶汽车和月球着陆等复杂问题。该系统将高度复杂的钻井环境建模为马尔可夫决策过程，然后在超级计算设施上进行训练，并使用现成的电子设备在实验室规模的钻井装置中进行验证。该项目将在钻井过程中实现自动化顺序决策过程。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Technology Transfer Program Phase I (STTR) project is to provide an environmentally friendly, sustainable, and cost-effective access to energy by improving information flow during drilling processes. The US geothermal energy industry is expected to grow by 48% to $6.8 billion by 2026 and will benefit from the significant savings of drilling activities, which currently constitute 40-60% of the costs of the entire geothermal plant and field development. In addition, oil and gas (O&G) drilling operations will benefit from faster and cleaner operations. This STTR Phase I project addresses the inability to transmit large amounts of data from downhole to the surface by performing most of the processing at the drill-bit. The proposed state-of-the-art drilling tool incorporates advanced processing power and intelligence downhole. It is built on physics-based machine learning algorithms centered on Reinforcement Learning (RL) techniques, used previously in complex problems such as autonomous vehicles and lunar landing. The system will model the highly complex drilling environment as a Markov Decision Process, then trained on supercomputing facilities and validated in a laboratory-scale drilling setup using off-the-shelf electronics. This project will lead to an automated sequential decision-making process in drilling.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.